What’s the problem?
Providing care for critically injured patients is challenging. Trauma centers, intensive care units, and emergency rooms confront these challenges amid noisy data and fast-paced, high-stress environments. This situation impacts the quality, efficiency, and consistency of care. Additionally, it increases the cognitive load on healthcare providers, leading to burnout and reduced productivity.
Who has the problem?
Health systems that serve critically injured and ill patients – usually the ones with an ACS designated trauma centers and Intensive Care Units. Specifically, administrative leaders over trauma, critical care, surgical care service lines will have the responsibility to improve quality and efficiency of care for these patients.
What’s the cost of not solving the problem?
It results in lower clinical outcomes for patients and lower operating margins for the health system. Lower clinical outcomes carry reputational risks for health systems and may also lead to losing accreditation from professional accreditation authorities (such as the ACS). Additionally, it can impact quality and safety ratings from CMS, resulting in lower payments.
From a financial perspective, critical care involves high fixed costs and unpredictable variable costs with capped per-patient revenue. The resulting thin operating margins for these patients are further squeezed for Medicaid patients or uninsured patients. Therefore, any reduction in hospital staff productivity and inefficiency in the utilization of scarce and expensive resources (such as beds, ventilators, and operating rooms) has a direct impact on the bottom line.
How is this currently solved? Why doesn’t that work?
Providers rely on physical evaluation and quantitative data—such as vitals, labs, and imaging data—to assess patients and apply the appropriate strategy for patient care. Currently, this quantitative data is made available to providers in its original form through Electronic Health Record (EHR) software. Several studies document the low usability of EHR software, which leads to excessive clicks and scrolling at the expense of clinical care. This problem is exacerbated in the fast-paced, large, multi-functional clinical team environment of critical care, where a flood of time-critical data is present. This situation introduces data fatigue and cognitive biases, leading to subjective and less-than-optimal patient care strategies.
What has changed?
AI techniques have vastly improved and have penetrated everyday usage and workflows. Consequently, clinicians are more inclined (and even expect) to adopt AI in their clinical workflows to better interpret quantitative data from patients. Requirements for data interoperability and standardization from CMS have led to widespread data availability, making it possible to train machine learning models specifically for critical care patients and deploy them in real time. The availability of HIPAA-compliant services on cloud platforms enables the cost-effective implementation and integration of EHR applications that embed AI-driven insights and analytics into critical care workflows.
How do you know it’s better?
End users in pilot programs have reported improvements in their decision-making and communication practices. We can also extrapolate from successes in other clinical domains where AI has been used to enhance decision-making. Our future pilots will continue with rigorous measurement and evaluation of clinical and financial metrics to quantify these improvements.